quantitative impact of the application of the matching
TRANSCRIPT
Masters Final Work Report
Internship at Autoridade de Supervisao de Seguros e
Fundos de Pensoes
Quantitative impact of theapplication of the Matching
Adjustment in the Portuguesemarket
Author:
Maria Ines Silva
Supervisors:
Dra. Ana Rita Ramos
Prof. Joao Andrade e Silva
Master in Actuarial Science
October 2015
INSTITUTO SUPERIOR DE ECONOMIA E GESTAO
Master in Actuarial Science
Quantitative impact of the application of the Matching Adjustment in
the Portuguese market
by Maria Ines Silva
Abstract
This report followed an internship at ASF, the Portuguese insurance supervisory
authority, during which I worked closely with the new insurance supervisory regime
(the Solvency II) and, in particular, the Matching Adjustment. This adjustment
aims to mitigate the impact of spread movements in assets backing specific obli-
gations. The main goal of this work was to quantify the impact of applying the
Matching Adjustment in Portuguese Worker’s Compensation annuities, through its
application in eligible portfolios (real or notional).
In short, we concluded that, although the market is not yet prepared to apply
the Matching Adjustment, in some cases, the adjustment can be advantageous to
undertakings. Also, the main impacts of the adjustment can be found in Technical
Provisions, where a decrease in the Best Estimate occurs due to the increase of the
risk-free rates, and in the Solvency Capital Requirement, where an increase of this
item is observed due to losses in diversification benefits. Lastly, by implementing
the entire process in Excel, we aimed to automate the computations related with
the Matching Adjustment and to make it easier to implement in the Portuguese
insurance market.
KEYWORDS: Solvency II, Long-term guarantees, Matching Adjustment, Work-
ers’ Compensation, Technical Provisions
INSTITUTO SUPERIOR DE ECONOMIA E GESTAO
Mestrado em Ciencias Actuariais
Impacto quantitativo do Ajustamento de Congruencia no mercado
segurador Portugues
por Maria Ines Silva
Sumario
O presente relatorio segue um estagio na ASF, a autoridade supervisora do mer-
cado segurador Portugues, durante o qual eu trabalhei com o novo regime europeu
de solvencia (o Solvencia II) e, em particular, com o Ajustamento de Congruencia.
Este ajustamento pretende mitigar o impacto de movimentos de spread em carteiras
de rendas dos ramos Vida e Nao Vida. O objectivo focal deste trabalho foi a quan-
tificacao do impacto do Ajustamento de Congruencia no mercado de seguros de
Acidentes de Trabalho em Portugal, atraves da sua aplicacao em carteiras elegıveis.
Resumidamente, foi possıvel concluir que, apesar de o mercado ainda nao estar
preparado para aplicar o ajustamento de congruencia, em alguns casos, o ajusta-
mento pode ser vantajoso para as seguradoras. Alem disso, os principais impactos
do ajustamento traduzem-se nas Provisoes Tecnicas, onde ocorre uma diminuicao
na Melhor Estimativa provocada por um aumento nas taxas de juros sem risco, e
no Requisito de Capital de Solvencia, onde ocorre um aumento deste item devido
a perda de benefıcios de diversificacao. Por fim, atraves da implementacao do pro-
cesso em Excel, objectivamos automatizar dos calculos relativos ao Ajustamento de
Congruencia e torna-lo mais acessıvel ao mercado segurador Portugues.
PALAVRAS-CHAVE: Solvencia II, Garantias de longo prazo, Ajustamento de
Congruencia, Acidentes de trabalho, Melhor Estimativa
Acknowledgements
First of all, I have to thank my supervisors, Prof. Joao Andrade e Silva and Dra. Ana
Rita Ramos, for all the discussions, insights and help given during my internship.
Also, I must recognize Prof. Hugo Borginho for believing in me and giving me this
amazing opportunity of working in ASF and developing this work. Then, I would
like to thank the entire DRS team and all the people at ASF that supported me and
made this work possible.
Secondly, I acknowledge my parents, Carlos Silva and Fatima Pastor, for all the
encouragement, friendship and support given throughout my life. Without their
teachings and their amazing role model, I wouldn’t be where I am and this work
wouldn’t be possible.
Finally, I thank all the friends and colleagues who accompanied me during my jour-
ney and shaped who I am today. In particular, Paulo Martins who supported me
during the last year and also made this work possible.
iii
Contents
Abstract i
Sumario ii
Acknowledgements iii
List of Figures vi
List of Tables vii
Abbreviations viii
1 Introduction 1
2 Theories, Regulations and Methodologies 3
2.1 Solvency II Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Matching Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 Workers’ Compensation . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Proposed problem and methodologies . . . . . . . . . . . . . . . . . . 10
3 Workers’ Compensation in the Portuguese insurance market 12
3.1 Liabilities portfolio . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Assets portfolio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
4 Matching Adjustment Application 19
4.1 Eligibility criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Building eligible assets portfolios . . . . . . . . . . . . . . . . . . . . 21
4.3 Matching Adjustment computation . . . . . . . . . . . . . . . . . . . 24
4.4 Remarks and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5 Quantitative Impact 29
5.1 Best Estimate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
iv
Contents v
5.2 Solvency Capital Requirement . . . . . . . . . . . . . . . . . . . . . . 31
5.2.1 Solvency Capital Requirement (SCR) in Matching AdjustmentPortfolios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.2.2 Approximation for an undertaking applying the Matching Ad-justment (MA) . . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.3 Final results and remarks . . . . . . . . . . . . . . . . . . . . . . . . 35
5.3.1 General impact of the Matching Adjustment . . . . . . . . . . 35
5.3.2 General impact of the Volatility Adjustment . . . . . . . . . . 37
6 Conclusion 39
A Technical Specification for the Liabilities Portfolio 42
A.1 Annual aggravation rates . . . . . . . . . . . . . . . . . . . . . . . . . 42
B Matching Portfolios 43
C SCR Standard Formula 47
D Excel and VBA sample functions 48
Bibliography 50
List of Figures
3.1 Liabilities yearly cash-flows in million euros . . . . . . . . . . . . . . 15
3.2 Assets yearly cash-flows in million euros . . . . . . . . . . . . . . . . 17
4.1 General distribution of the MA, in basis points, per group . . . . . . 26
vi
List of Tables
4.1 Eligibility criteria for the Matching Adjustment’s application . . . . . 20
4.2 Information on Matching portfolios’ general features . . . . . . . . . . 23
4.3 MA and relevant spreads for its computation, in basis points (1 basispoint equals 0.0001) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.4 MA distribution’s general features . . . . . . . . . . . . . . . . . . . . 27
4.5 MA distribution and relevant spreads, in basis points . . . . . . . . . 27
5.1 Quantitative impact in the Best Estimate . . . . . . . . . . . . . . . . 30
5.2 SCR for MA Portfolios and corresponding baselines . . . . . . . . . . 33
5.3 Relative impact on solvency of the Matching Adjustment . . . . . . . 36
5.4 Relative impact on solvency of the Volatility Adjustment . . . . . . . 38
A.1 Aggravation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
vii
Abbreviations
AAER Assets Annual Effective Rate
ASF ”Autoridade de Supervisao de Seguros e Fundos de Pensoes”
BE Best Estimate
BOF Basic Own Funds
CQS Credit Quality Step
D&D Death and Disability
EIOPA European Insurance and Occupational Pensions Authority
FS Fundamental Spread
LAER Liabilities Annual Effective Rate
LTG Long-Term Guarantee
LTGA Long-Term Guarantee Assessment
MA Matching Adjustment
QIS Quantitative Impact Study
SCR Solvency Capital Requirement
SLT similar-to-life techniques
viii
Abbreviations ix
VA Volatility Adjustment
VBA Visual Basic for Applications
WC Workers’ Compensation
Chapter 1
Introduction
This report follows an internship at ”Autoridade de Supervisao de Seguros e Fun-
dos de Pensoes” (ASF), the Portuguese insurance supervisory authority. During 6
months, I worked with the Risk and Solvency Analysis team, called DRS. The team
is responsible for the analysis and reporting on the solvency and financial state of the
Portuguese insurance and pension funds market and for the institutional representa-
tion, which includes downscaling the European laws and guidelines, working within
the committees and working groups of the European Insurance and Occupational
Pensions Authority (EIOPA) and supporting the supervision teams by answering
questions within its scope of knowledge.
Besides the main research work I developed, which was the aim of the internship
and will be the focus of the report, I also integrated other daily activities of DRS,
including the drafting and data analysis of the annual report on the Portuguese
insurance and pension funds activity and the drafting of regulatory standards. It was
very rewarding because it helped me understand the team’s usual activities, apply
the knowledge I acquired during my masters and learn more about insurance and
finance. Due to the nature of DRS’s role in ASF, the team works very closely with
Solvency II, the new European solvency regime which will be discussed in Chapter
2. Being a new regulation which has not entered into force yet, there’s still a lot to
1
Chapter 1.Introduction 2
be done to fully understand its future consequences, its advantages/disadvantages
and the regime itself. This fact is specially relevant for the Long-term Guarantee
package, which was added by an amendment to the regime approved in 2014 and
is the motivation for my internship. Indeed, I was asked to develop a study on the
quantitative impact of applying the Matching Adjustment (a measure within the
Long-term Guarantee package) to the Portuguese Workers’ Compensation insurance
market and the main goals given to me were the following:
1. Do a brief overview of the subject, including the Solvency II regime and the
Matching Adjustment;
2. Analyze the requirements to apply the Matching Adjustment and to select
eligible portfolios which are representative of the Portuguese Workers’ Com-
pensation insurance market;
3. Measure the impact of applying the Matching Adjustment to real or notional
portfolios, namely in Technical Provisions, Own Funds and Solvency Capital
Requirements;
4. Extend the analysis to other measures in the Long-term Guarantee package,
in particular, the Volatility Adjustment.
Thus, Chapter 2 includes a description of Solvency II, the Matching Adjustment
and the Long-term Guarantee package, the Workers’ Compensation insurance in
Portugal and the general methodologies used to deal with the proposed problem.
Chapter 3 presents a general analysis about the state of the Workers’ Compensation
insurance in Portugal during 2014. Chapter 4 explores the possibility of applying
the Matching Adjustment in Portugal, including a revision of the eligibility criteria,
the construction of eligible portfolios representative of market and the computation
of the measure itself. Chapter 5 includes the valuation of the Best Estimate (a part
of Technical Provisions), the computation of the Solvency Capital Requirement and
the assessment of the general impact of applying the Matching Adjustment. Finally,
Chapter 6 concludes the report by summarizing the main findings and setting the
path for future developments.
Chapter 2
Theories, Regulations and
Methodologies
2.1 Solvency II Overview
Since the 1970s, the European Union has been aiming for a harmonization of prac-
tices and a convergence of supervision for the insurance sector within the European
Economic Area. That was the decade when the first directives were implemented
and since then we have been walking towards that goal. Many revisions have been
made, however they were unable to accompany the rapid changes occurring in the
european insurance sector and the financial markets. Thus, the current framework,
named Solvency I and regulated by the directives (Council, 1973) and (European
Parliament and Council, 2002), although easily implemented and providing a rel-
ative protection to policyholders, remains short on some aspects. The problem is
that these directives are insufficient in showing an insurer’s true risk exposures, lead
to a misalignment between capital requirements and the risks the insurer is exposed
to, do not fully incentive insurers to have a risk-based management and finally do
not provide sufficient tools for the supervisors to intervene.
3
Chapter 2.Theories, Regulations and Methodologies 4
Recognizing the need for a deep review of Solvency I, in 2001 began the development
of the project Solvency II, which aimed for a completely new regime instead of a
building-up on top of Solvency I. That was the kick-off year of many studies aimed
to develop a framework for the new solvency regime, being the Sharma report one of
them (Sharma, 2002). The main conclusion here was to base the new regime on Basel
II, a set of international regulations for finance and banking which is also structured
on three-pillars combining qualitative and quantitative measures. Then, following
the Lamfalussy Process (well described by Raptis (2012)), the Solvency II Frame-
work Directive was adopted (see European Parliament and Council, 2009). However,
after the 2008 financial crisis, some issues about the EU supervisory structure and
the impact of short-term market movements in long-tailed businesses arose, bringing
a lot of debate and delaying the implementation of Solvency II. Some studies and
reviews were made and finally, in 2014, the Omnibus II Directive (European Par-
liament and Council, 2014), amending the initial directive, and the Delegated Acts
(European Commission, 2014), describing implementing measures, were approved.
Presently, EIOPA is concluding its work on Technical Standards and Guidelines to
ensure consistent implementation and cooperation within the EU. The process will
be finished in 2016, time when the new regime will come into force.
Solvency II includes 2 important features - a global and integrated view of risks
(with the 3 main pillars) and the principle of proportionality, which sets that the
requirements should be proportionate to the risks’ nature, size and complexity.
The 3 pillars where the framework is built upon cover quantitative requirements (Pil-
lar 1), qualitative requirements (Pillar 2) and reporting and disclosure requirements
(Pillar 3). Together, they give a complete risk-based overview of insurance compa-
nies, always encouraging undertakings to manage their risks and being transparent.
The quantitative requirements in Pillar 1 contain the following main components:
Integrated view of the balance sheet - Both assets and liabilities are valued
through market consistent principles in a total balance sheet approach, which
Chapter 2.Theories, Regulations and Methodologies 5
means that not only all assets and liabilities are valued, but also their inter-
actions are considered in the valuation.
Technical provisions - They represent the amount necessary to cover all insur-
ance liabilities in force at the valuation date and it’s subdivided into the Best
Estimate (present value of the liabilities cash-flows’ expected value, discounted
using the risk-free rates) and the risk margin (cost of holding regulatory capital
sufficient to ensure the full run-off of liabilities).
Own funds - Financial resources available to create new business and to absorb
unexpected losses. They’re divided into Basic Own Funds (the excess of assets
over liabilities plus subordinated liabilities) and Ancillary Own Funds (off-
balance-sheet commitments that the insurer may call to increase its financial
resources). They’re classified in 3 tiers based on their loss-absorbing capacity
and subordination.
Capital requirements - Represents the extra capital needed to sustain unexpected
losses and has two levels: the Solvency Capital Requirement - the capital
needed to sustain a shock corresponding to the value-at-risk 99.5% for one
year time horizon - and the Minimum Capital Requirement - level from where
the risk of insolvency is considered excessive. Both capital requirements must
be covered by specific Own Funds items and any breach implies actions from
the supervisor.
Despite being in its implementing stages and still having some areas in development,
Solvency II is a very ambitious project. By being risk-based, taking an economic view
and promoting group supervision, the new regime will encourage financial stability,
a higher protection of policyholders and a more efficient operation of the insurance
sector, aiming thus to deepen the current single market across the EU.
Chapter 2.Theories, Regulations and Methodologies 6
2.2 Matching Adjustment
As discussed before, one of the reasons why the implementation of Solvency II was
delayed was the intense debate around the valuation of long-term liabilities and the
long-term guarantees problem. Because Solvency II is risk-based and sensitive to
market conditions, it was failing to address situations where insurance products with
long-term guarantees were being affected by artificial volatility and pro-cyclicality.
To solve this, during 2013, EIOPA carried out a Quantitative Impact Study (QIS)
on long-term guarantees, the Long-Term Guarantee Assessment (LTGA). The study
tested 6 different measures feasible to be included in the new framework, namely
the Counter-Cyclical Premium, some versions of the Matching Adjustment, extrap-
olation methods and the extension of recovery period. The goal was to quantify the
impact of these measures in the valuation of liabilities and the calculation of capital
requirements. Some measures were rejected while others suffered some modifications
and finally, in 2014, Omnibus II adopted the definitive Long-Term Guarantee (LTG)
package within the Pillar 1 requirements, which includes now the current Matching
Adjustment (MA).
The MA was especially built for life and non-life annuity products, aiming to mitigate
the impact of spread movements in the value of Technical Provisions and capital
requirements for some special portfolios. In fact, because those products’ liabilities
are predictable and can be fully matched by assets held until maturity, undertakings
are not forced to sell their assets to meet their liabilities and thus they are not
exposed to short-term market volatility. Also, because the MA is tailor-made for each
undertaking, even in its early stages, the LTGA report acknowledged the Matching
Adjustment as ”the most effective tool within the LTG package with regards to
mitigating short-term volatility from Solvency II balance sheets of portfolios eligible
for the measure” (EIOPA, 2014).
Chapter 2.Theories, Regulations and Methodologies 7
Any undertaking wishing to apply this measure must comply with several eligibility
criteria addressing the portfolio of assets, the portfolio of liabilities and the manage-
ment of both portfolios. Summarily, their purpose is to ensure that:
• Both portfolios are managed separately from the rest of the business;
• Liabilities do not give rise to future premiums and allow no options for pol-
icyholders, except for surrender options where the surrender value does not
exceed the value of the assets;
• The only underwriting risks to which liabilities are exposed to are expense,
revision, longevity and immaterial mortality;
• The asset portfolio is composed of bonds or similar cash-flows instruments
which are held until maturity and have fixed cash-flows;
• The cash-flows of the asset portfolio replicate the liabilities cash-flows and any
mismatch doesn’t give rise to material risks.
Since the asset portfolio is only exposed to the risks of downgrade and default and
since the eligible liabilities are fully covered by that portfolio, the undertaking is
given the benefit of discounting those liabilities at a higher rate than the risk-free
rate. This higher rate is given by the risk-free rate plus a flat adjustment (the so-
called MA) based on the asset portfolio’s spread adjusted to account for the two risks
it is in fact exposed to. In other words, the MA is a parallel shift of the risk-free
interest rate term structure used to discount the Best Estimate (BE) of the portfolio
of eligible liabilities and it is equal to the asset portfolio market yield minus the risk-
free rate and the Fundamental Spread (FS). This FS reflects the assets’ probability
of default plus the expected loss due to downgrade and it’s computed quarterly by
EIOPA, using long-term statistics. For more details regarding the computation of all
needed statistics by EIOPA can be found in the Technical Specifications by EIOPA
(2015).
Lastly, an undertaking can only apply this measure under approval by the super-
visory authority, following a specific application process. The application should
Chapter 2.Theories, Regulations and Methodologies 8
contain the methods used by the undertaking to manage both portfolios and to ap-
ply the measure. Once approved, the undertaking cannot revert back to an approach
where the MA is not applied, even in cases of extremely low spreads when the MA
could become negative. Also, the scope of such approval may cover all future item
in the matching portfolio, provided the undertaking can prove that the matching
portfolio continues to meet all relevant requirements and that those new items have
the same features as the current matching portfolio.
2.3 Workers’ Compensation
Workers’ Compensation (WC) is a program which provides medical benefits and/or
income replacement in case of occupational illnesses and injuries. In 1884, Germany
was the first country to introduce a WC program. After that, many countries
adopted it, although each one with its own particularities.
WC is divided in 3 main program types. WC programs can be integrated within
a wider social insurance program, also providing, for instance, retirement pensions.
They can be within a WC compulsory insurance system, where employers are liable
to insure its WC liabilities. And they can be within a voluntary WC insurance
system, where employers are responsible for all WC obligations but are not obliged
to secure those obligations via insurance. Portugal can be considered to have a
mixed security system, since there’s a clear division between occupational accidents
and occupational illness. While for the first, it is a compulsory insurance system,
for the second, the responsibility falls in the public security system. Thus, when
working within WC insurance, undertakings in Portugal only deal with occupational
accidents.
The Portuguese law defines an occupational accident as any sudden and unforeseen
event which occurs in the workplace and during working time, that causes, directly
or indirectly, bodily injury, functional disorder or illness and results in a reduction
Chapter 2.Theories, Regulations and Methodologies 9
in the person’s earning/working capabilities or in death. The sudden nature of
accidents is the main difference between them and occupational illnesses, which
arise from a long and continued exposure to a certain occupational risk.
The WC insurance for occupational accidents provides two main groups of benefits
- benefits in kind and benefits in cash. The first includes essentially medical treat-
ments, while the second mainly relates to income replacement. Contained in the
cash benefits are the Death and Disability (D&D) pensions. In Portugal, disability
pensions are granted to victims who have already been submitted to all relevant
treatments and whose disability is considered permanent. This doesn’t mean that
the disability won’t change. In fact, it’s not uncommon to see the medical con-
dition evolve either for better or worse and, in these cases, both beneficiary and
undertaking may ask for exactly one legal disability revision per year. Disability
pensions are perpetual and are computed according with the victim’s yearly income
and disability, which means that disability revisions lead to changes in pension val-
ues. Finally, disability pensions undergo three legal states - provisional, defined and
definitive. In the first, there’s no legal agreement on the disability and yearly in-
come, in the second, the disability is legally established, and in the third everything
is legally established and the pension’s value is fully determined. With respect to
the death pensions, specific types of beneficiaries receive fixed percentages of the
victims’ yearly income. Usually, these pensions are fixed and perpetual, but there
are some exceptions. For instance, orphans receive pensions only until they’re 18
years old or until they are 22 or 25 and still attending school, while spouses receive
an extra 10% when they complete 65 years of age.
There’s also a public fund for WC in Portugal, the ”Fundo de Acidentes de Tra-
balho”. It is managed by ASF and has two main roles - it is responsible for covering
obligations when the entity responsible for those obligations cannot meet them (for
instance, in case of insolvency) and it reimburses undertakings for the yearly in-
flation updates on pensions’ values (this way, undertakings only pay fixed pension
amounts).
Chapter 2.Theories, Regulations and Methodologies 10
More information on WC in Portugal can be consulted in Alegre (2011), which
includes all the relevant Portuguese laws explained. In conclusion, although the MA
was not originally designed for the WC insurance business, given the particularities
of this line of business in Portugal, not only D&D pensions are well suited for the
measure’s application, but also it may be the line of business where the measure will
have the most relevant impact.
2.4 Proposed problem and methodologies
As discussed before, the internship aimed to study the quantitative impact of apply-
ing the Matching Adjustment to the Portuguese WC insurance market. Therefore,
the following plan was devised:
1. Cluster all 15 undertakings operating in the Portuguese WC sector into three
groups according with their market shares. They are the Big Dimension (BD)
group, the Medium Dimension (MD) group and the Small Dimension (SM)
group. They represent 81%, 17% and 2% of the market, respectively, and
include 6,5 and 4 undertakings, respectively.
2. Compute the yearly expected liabilities cash-flows, related with each group’s
current D&D pensions.
3. Study the composition of each group’s assets portfolio covering the respective
liabilities.
4. Check the requirements to apply the MA for each group.
5. Find assets portfolios that fully match each group’s liabilities and are repre-
sentative of the groups’ original assets portfolio. They’ll be called Matching
portfolios.
6. Compute the MA for each set of portfolios.
7. Compute the Best Estimate for each group’s portfolios, with and without MA,
and assess its impact.
Chapter 2.Theories, Regulations and Methodologies 11
8. Compute the Solvency Capital Requirement (SCR) for each group’s portfolios,
estimate the SCR such undertakings would have, with and without MA, and
assess its impact.
In order to make steps 5 to 8 as faster to run and as automated as possible, we
implemented them in Excel workbooks which included functions programmed in
Visual Basic for Applications (VBA). Appendix D contains parts of the workbooks
and VBA’s coding. The full content can be available upon request.
In the end, we hope to evaluate the full impact of applying the MA to three no-
tional undertakings - the BD, MD and SM groups, which are representative of three
different types of Portuguese insurers.
Chapter 3
Workers’ Compensation in the
Portuguese insurance market
Chapter 3 analyses the current state of WC insurance in Portugal, bearing in mind
it only includes occupational accidents. Also, since the MA is specially built for
annuities, we just considered D&D pensions. In order to perform this analysis, we
used two sets of data, namely:
Liabilities portfolio - Information reported to ASF on all current beneficiaries,
according with the regulatory standard issued by Instituto de Seguros de Por-
tugal (2007).
Assets portfolio - Details concerning the assets belonging to the assets portfolio
covering all WC liabilities, which is called ”Carteira 5” in Portugal. The
aforementioned data is annually reported to ASF and was backed by data
provided by Bloomberg.
All data refers to the year 2014, although we also used information on the liabilities
portfolio for the years 2012 and 2013 to estimate some parameters. As explained
in the previous chapter, the cash-flows of each specific undertaking were aggregated
through their market shares’ weighted average.
12
Chapter 3.Workers’ Compensation in the Portuguese insurance market 13
3.1 Liabilities portfolio
The main goal of this analysis was to project the full run-off of all pensions’ cash-
flows. For each beneficiary, we had information on the insurance company, gender,
actuarial age, type of beneficiary, pension value, pension state and existence and
reason for revisions of the pension value. Thus, assuming that pensions are paid at
the beginning of the year, we built an iterative formula to compute the basic pension
value P yx for all beneficiaries aged (x) during the year y:
P yx = P y−1
x−1 · px−1 for y > 2015 and P 2015x = P initial
x (3.1)
where px is the one-year survival function for a person aged (x) and P initialx is the
sum of all pensions paid during the year 2014 for the beneficiaries aged (x) and
alive by the end of 2014. To compute px, we chose the specific mortality tables used
by each undertaking for computing their current provisions and, for simplicity, we
assumed that pensions were paid at the beginning of the year.
Besides the general formula, we had to take into account some specificities of the
D&D pensions:
• The beneficiaries had to be divided by gender because the mortality tables
differ from male to female.
• Since, for the purpose of applying the MA, we are only interested in annuities,
we excluded all redeemable (i.e. paid as a lump sum), suspended and extinct
pensions, leaving only the active and non redeemable pensions.
• As seen in Chapter 2, different types of beneficiaries have different rules con-
cerning the pensions’ run-off and so we divided all types into 4 groups and
adjusted the general formula to (P yx )∗:
General annuity - these are the plain vanilla pensions, i.e., constant pen-
sions granted until the beneficiaries’ death. The formula is the same
(P yx )∗ = P y
x .
Chapter 3.Workers’ Compensation in the Portuguese insurance market 14
+5% annuity - also perpetual but when the beneficiary completes 65 years
of age, the pension is increased by 5%. The pension is given by the general
formula until age 65 and after that is equal to (P yx )∗ = P y
x · 1.05.
+10% annuity - the same as the +5% annuity, but has an increase of 10%.
After 65 years of age, the pension is (P yx )∗ = P y
x · 1.10.
Orphan annuity - term annuity that stops when the beneficiary completes
25 years of age. The pension is the same (P yx )∗ = P y
x until 25 years of age
and after that is zero.
• Disability pensions suffer changes in their values due to revisions of the disabil-
ity attributed to the beneficiary and due to changes of state from provisional
to definitive. To include this in the projections, we computed the annual ag-
gravation rates for 2012, 2013 and 2014 (using the data for these years) and,
for prudence, chose the highest rate - 4.52% (the results are in appendix A,
Table A.1). Thus, the formula to compute pensions’ value was again adjusted
to (P yx )∗∗ = (P y
x )∗ · 1.0452.
• Based on the available data, it was possible to conclude that the amount of
expenses related to D&D pensions were immaterial, and thus we didn’t include
expenses in the cash-flows’ projection.
After obtaining the annual pensions for each actuarial age (x), we added these
pensions to find the total nominal pension amount for each year. Then, for each
year, we aggregate these cash-flows by computing their weighted average according
with the undertakings’ market shares. The results are displayed in Figure 3.1. As
expected, the BD group has a significantly higher amount of yearly liabilities cash-
flows, when compared with the other two groups. Also, its liabilities are more
condensed in earlier maturities than the MD group. In fact, the BD group has its
peak of cash-flows around the year 2035, while the MD group has its peak around
2047 and presents smaller relative values of cash-flows in the first years. Although it
is much smaller, the SD group presents a structure of its liabilities yearly cash-flows
more similar to the BD group’s.
Chapter 3.Workers’ Compensation in the Portuguese insurance market 15
Figure 3.1: Liabilities yearly cash-flows in million euros
These are the cash-flows which were later used to build Matching Portfolios and to
compute Technical Provisions and SCR.
3.2 Assets portfolio
There were two main goals for this analysis. Firstly, we wanted to understand
the portfolio’s composition and main features and, secondly, we wanted to know
if the assets matched the liabilities, i.e., if, at any given year, the nominal value
of liabilities was equal to the nominal value of assets, which is a requirement for
the MA’s application. From the first part of the analysis, we drew the following
conclusions:
General portfolio composition - All portfolios are majorly composed of sovereign
and corporate debt (bonds), adding always up to more than 60% in all the 3
groups. The BD group is the most diverse with material investment in equity,
property, real estate investment trusts and deposits. It is followed by the SD
group, with material investment also in property. Finally, the MD group is
the less diverse without material investment other that sovereign and corporate
debt.
Chapter 3.Workers’ Compensation in the Portuguese insurance market 16
Bonds’ coupon type - The most common coupon type is the fixed coupon (all
groups have more than 90% of this type), which pays a fixed coupon amount.
The other two material types are the floating coupon, where the coupon de-
pends on a money-market index, and the variable coupon, where the coupon
depends on other indexes or specified formulas which are based on market
conditions.
Bonds’ maturity type - The BD group is again the more diverse, with 83% of
the ”at maturity” type (the plain vanilla bonds) and both callable (the issuer
has the option of early redemption) and puttable (the holder has the option
of early redemption) types still material but below 6%. The MD group has
99% of ”at maturity” bonds and the SD group has 95% of ”at maturity” and
a material amount of callable bonds.
Bonds’ actuarial basis - This is the day count convention used to accrue interest.
The most common is the ACT/ACT with all groups with more than 88%. Both
ISMA-30/360 and ACT/360 are material.
Bonds’ Credit Quality Step (CQS) - This is a Solvency II convention that char-
acterizes the credit worthiness of assets. It is an integer scale from 0 to 6 where
the lower the CQS, the higher the rating. More information on this mapping
can be found at (Joint Committee, 2015). The BD group has an average CQS
of 4.10, the MD group has 2.86 and the SD group has 3.37.
Bonds’ yield - The BD group has an average bond yield of 1.9930, the MD has
1.4114 and the SD has 1.1927. For this analysis, for obvious reasons, we
excluded bonds with variable and floating coupons.
MA eligibility - Within ”Carteira 5”, the assets eligible for the MA’s application
are bonds (Corporate and Sovereign) with fixed cash-flows (mainly fixed and
zero-coupon) which allow no options for issuers. In the BD group these as-
sets correspond to 57.67% of the total portfolio, while in the MD group they
correspond to 91.64% and in the SD group correspond to 62.75%.
Then, for the second part of this study, we had to project the cash-flows of all
Chapter 3.Workers’ Compensation in the Portuguese insurance market 17
eligible assets. Knowing the bonds’ coupon yield, coupon type, frequency, maturity
date and redemption amount, we built the payoff matrix C ∈ Mn×m, where each
row corresponds to the cash-flows of a specific bond and each column corresponds
to the cash-flows of a specific year. We made some assumptions for simplicity:
• To project the liabilities, we assumed that they were paid in the beginning of
each year. Therefore, in order to be coherent, we discounted all cash-flows of
a certain year to the beginning of that year using the forward rates related to
the risk-free interest rates published by EIOPA. In this way, the payoff matrix
represented the value of assets’ cash-flows at the beginning of each year.
• Since the ACT/ACT was the actual basis more frequent, we used it for all
bonds.
With the payoff matrix C and the column vector w containing the number of units
invested in each bond1, the annual assets cash-flows were simply the row vector Cw.
The results are shown in Figure 3.2.
Figure 3.2: Assets yearly cash-flows in million euros
Once again, the BD group presents higher values in its cash-flows than the other
groups. However, the gap between them is lower in this case. Also, assets’ cash-
flows are much more irregular and concentrated in earlier years than the liabilities’
cash-flows.
1Information included in the data reported to ASF
Chapter 3.Workers’ Compensation in the Portuguese insurance market 18
A more attentive eye may notice that the assets’ cash-flows are much higher than
the liabilities’ cash-flows. The reason behind this considerable mismatch is the fact
that, even though for the purpose of applying the MA we are only considering active
and non redeemable pensions, the portfolio ”Carteira 5” is used to cover all future
WC obligations, including redeemable pensions, medical expenses, claims reported
but not settled and claims incurred but not reported.
Chapter 4
Matching Adjustment Application
4.1 Eligibility criteria
In order to apply the MA to specific portfolios of assets and liabilities, they need to
comply a set of criteria defined in the amendment made to the Solvency II Directive
by the European Parliament and Council (2014), the Omnibus II Directive. For each
group of undertakings (BD, MD and SM), we checked the fulfillment of all criteria
and, in the end, all groups failed the same requirements. Indeed, this outcome was
expected as undertakings have no reason to adapt their asset portfolios to comply
with such criteria in the current regime. The results for all groups are summarized
in Table 4.1.
In general, all criterion related with the liabilities portfolio is satisfied because the
nature of the D&D pensions’ business in Portugal fits perfectly within the scope
of the MA. It includes non-life annuities only exposed to the underwriting risks of
longevity, expenses and revision, and such that no future premium payments are
generated by them. However, the same does not happen with the assets portfolio,
since it fails every criteria. First of all, the portfolio ”Carteira 5” is used to cover all
WC liabilities, which disagrees with the requirement setting that assets cannot be
19
Chapter 4.Matching Adjustment Application 20
used to cover losses from other activities. Also the aforementioned portfolio is very
diversified, including investment in stocks and properties, which are asset types not
allowed for the MA’s application. Then, considering only the investment on bonds,
there are material investments in floating and variable coupon bonds, which fails the
criteria requiring assets with fixed cash-flows. Finally, if we only used the eligible
types of assets in the assets portfolio, it would not replicate the liabilities’ cash-flows
because there’s a huge gap between the maturities of the two portfolios. Actually,
all assets mature before 2044, while liabilities ultimately run past the year 2100.
Criteria Compliance
Assets portfolio
Consists of bonds and other assets with similar cash-flow characteristics No
Is maintained over the lifetime of obligations No
Cannot be used to cover losses arising from other activities No
Its expected cash-flows replicate the expected cash-flows of liabilitiesand any mismatch does not give rise to material risks
No
Its cash-flows are fixed and cannot be changed by issuers or third parties(a dependence on inflation is allowed provided that those assets are usedto cover liabilities that depend on inflation)
No
Liabilities portfolio
Includes life insurance obligations and/or non-life annuities Yes
All contracts underlying this portfolio cannot give rise to future premiumpayments
Yes
The only underwriting risks connected with this portfolio are longevity,expense, revision and mortality
Yes
When there’s an exposure to mortality risk, the BE cannot increase bymore than 5% under a specified mortality risk stress
NotApplicable
Includes no options for policyholders except a surrender option wherethe surrender value does not exceed the value of assets
Yes
Obligations of a single contract cannot be split into different partswhen composing this portfolio
Yes
Both portfolios
Are identified, organized and managed separately from otheractivities of the undertaking
No
Table 4.1: Eligibility criteria for the Matching Adjustment’s application
Chapter 4.Matching Adjustment Application 21
Therefore, we conclude that the market is not yet prepared to apply this measure.
Nevertheless, we still want to estimate what would be the MA’s impact in the
Portuguese market, assuming it could indeed be applied. Since the problem resides
in the assets portfolio, we can build eligible assets portfolios for the market’s real
liabilities portfolios and continue the analysis from there.
4.2 Building eligible assets portfolios
The process of building eligible portfolios of assets consisted of three main steps:
1. Find a set of eligible assets with maturities ranging from 2015 to 2074;
2. Select one asset for each different maturity year depending on specified criteria;
3. For each asset, compute the amount of units that should be held in order to
replicate the liabilities’ cash-flows.
Our aim was to build portfolios as close as possible to the real portfolios the un-
dertakings already held and to use always real assets being traded in the market by
the end of 2014. Thus, for step one, we gathered all eligible assets from the original
portfolios and, for maturities not present in those portfolios, we added all eligible
assets quoted in Bloomberg and exchanged in euro currency. In the end, we had 3
sets of eligible assets, one for each specific group of undertakings.
For step two, we considered two distinct investment strategies - the portfolio with
the highest yield and the portfolio with the highest rating. The first is quite obvious
- for each maturity we chose the assets with the highest yield to maturity computed
from the 31st of December, 2014. For the second, we didn’t use the assets’ rating,
but instead their CQS. For each maturity, we chose the asset with the lowest CQS
and, to break a tie, we chose the asset with the highest yield.
Then, for step 3, we used the following method:
Chapter 4.Matching Adjustment Application 22
Consider n different bonds, with market prices B1, B2, ..., Bn, and their payoff matrix
defined as C = [cij]n×n, where cij is the cash-flow of bond i at time tj. Consider also
the liabilities vector L = [lj]n×1 where lj is the expected liabilities cash-flow at time
tj. Now, for the portfolio w = [wi]n×1, where wi is the number of bonds i held in it,
we have that the vector of the assets cash-flows is given by wTC.
The condition is both liabilities and assets having the same cash-flows at all times.
Thus w has to be such that wTC = LT , which is equivalent to wT = LTC−1 if
and only if C is invertible. Since both investment strategies define portfolios with
exactly one asset per maturity year, from 2015 to 2074, the payoff matrix C is
lower triangular, which implies that C is invertible. Therefore, wT = LTC−1 is the
portfolio that perfectly matches L.
However, this method has some drawbacks and thus some adjustments were made:
• As explained in chapter 3, liabilities’ cash-flows were computed assuming all
pensions were paid at the beginning of the year. However bonds can mature at
any time during the year. Since we need the assets first to be able to pay the
liabilities, we considered that any bond maturing in a certain year would be
used to cover the cash-flows of the next year. For instance, all bonds maturing
in 2015 will be used to cover the liabilities of 2016.
• Although liabilities run past the maturity year 2100, the current market does
not contain eligible assets with so long maturities. Therefore, to solve this
issue, we considered that the highest maturity year would be 2075 and the
assets maturing in that year would have to cover all future liabilities cash-
flows. In other words, we added the nominal value of all cash-flows being paid
after the year 2075 and assigned that value to the year 2075.
• Although one can only hold an integer number of bonds, this method may lead
to non integer values for wi. In these cases, we rounded the value obtained to
the nearest integer. Despite the approximation provoking a certain degree of
Chapter 4.Matching Adjustment Application 23
mismatch between the assets and liabilities cash-flows, the mismatch was not
material.
• The method may also lead to negative values of wi, which is not acceptable.
In these cases, we considered wi to be zero. Again, the approximation leads to
mismatches. However, the only portfolios where we had to apply it were the
two covering the MD group and even in those the mismatch was not material.
Finally, when valuing assets’ cash-flows for applying and computing the MA, the
Directive states that those cash-flows must be adjusted to take into account the risk
of default. In particular, cij = cij · (1 − PDi) + 0, 3 · cij · PDi1, where PDi is the
probability of default for bond i, which is provided by EIOPA and depends on the
bond’s CQS, maturity and type of issuer (i.e., whether it is Governmental, Financial
or Non-financial). Thus, the payoff matrix C was adjusted in step 3 according to
the presented formula.
In the end, after implementing the entire process in Excel using functions pro-
grammed in VBA, we built 6 different eligible portfolios (two for each group of
undertakings), that match the correspondent cash-flows of liabilities. The full con-
tent of these Matching portfolios is in Appendix B and their general features are
presented in Table 4.2.
Group PortfolioAverage
YieldAverage
CQSAverageDuration
% Invested inGovernment
% ofMismatch
BD H. Yield 0.0365 4.40 18.62 51.12% 0.12%H. Rating 0.0194 1.70 19.91 47.03% 0.04%
MD H. Yield 0.0332 4.32 24.76 46.26% 0.36%H. Rating 0.0186 1.70 25.35 45.37% 0.58%
SD H. Yield 0.0352 4.10 18.24 37.05% 3.37%H. Rating 0.0168 1.16 19.54 37.98% 2.72%
Table 4.2: Information on Matching portfolios’ general features
1the formula is given in EIOPA (2015)
Chapter 4.Matching Adjustment Application 24
4.3 Matching Adjustment computation
In the Directive’s amendment by European Parliament and Council (2014), the
Omnibus II Directive, the MA is defined to be the difference between two annual
effective rates, the Assets Annual Effective Rate (AAER) and the Liabilities An-
nual Effective Rate (LAER), minus the portfolio’s Fundamental Spread (FS). The
directive outlines these concepts as:
AAER - The annual effective rate, calculated as the single discount rate that,
where applied to the cash-flows of the portfolio of insurance or reinsurance
obligations, results in a value that is equal to the market value of the portfolio
of assigned assets.
LAER - The annual effective rate, calculated as the single discount rate that, where
applied to the cash flows of the portfolio of insurance or reinsurance obligations,
results in a value that is equal to the value of the BE of the portfolio of
insurance or reinsurance obligations.
Fundamental Spread- the sum of two credit spreads, the one corresponding to
the probability of default of assets and the one corresponding to the expected
loss resulting from downgrading of the assets. Both spreads and the FS are
provided by EIOPA and depend on the assets’ CQS, maturity and type of
issuer. Moreover, the FS used to subtract from the MA must include only the
portion of the FS that has not already been reflected in the adjustment to the
cash-flows of the assigned portfolio of assets. And finally, the FS of each asset
must be at least 30% (for exposures to Member States’ central governments
and central banks) or 35% (for any other exposures) of the correspondent long-
term average of spread over the risk-free interest rate. These long-term average
of spread is also computed by EIOPA periodically.
In practice, we computed the MA by applying the following 7 steps in Excel, with
the support of functions programmed in VBA:
Chapter 4.Matching Adjustment Application 25
1. Compute the market value of the Matching portfolio as the sum over all assets
of their real market value multiplied by the number of units invested in them,
wi;
2. Compute the BE of the liabilities portfolio as the present value of all liabilities’
cash-flows using the risk-free rate term-structure published by EIOPA;
3. Compute both AAER and LAER, as described in the directive, using the
Excel’s add-in named Solver;
4. For each asset, get its FS and credit spread corresponding to the probability
of default of assets and compute its remaining FS as the difference of the two;
5. Compute each asset’s duration;
6. Compute the weighted average of the assets’ remaining FS, using the amounts
invested in each asset and their duration;
7. Finally, compute the MA as the AAER minus the LAER and minus the
weighted average remaining FS.
4.4 Remarks and Results
Group Portfolio AAER - LAER - FS = MA
BD H. Yield 274.08 - 170.13 - 2.68 = 101.27H. Rating 173.20 - 170.13 - 1.28 = 1.79
MD H. Yield 251.11 - 196.63 - 3.40 = 51.08H. Rating 164.31 - 196.63 - 2.01 = -34.34
SD H. Yield 274.91 - 166.03 - 2.66 = 106.23H. Rating 153.58 - 166.03 - 1.06 = -13.51
Table 4.3: MA and relevant spreads for its computation, in basis points (1 basispoint equals 0.0001)
First of all, it is clear that the Highest Yield portfolios have consistently higher MA’s
than the Highest Rating portfolios. This trend seems to be caused by the influence
the yields have on AAER’s and the fact that lower ratings don’t influence in the
same degree the FS.
Chapter 4.Matching Adjustment Application 26
Also, there are evident differences between the three groups of undertakings. Firstly,
the MD group seems to have a generally lower MA than the other groups, not only
due to higher FS’s, but also due to higher LAER’s and, in some cases, lower AAER’s.
Secondly, the SD group presents a higher variation between the Highest Yield and
Highest Rating portfolios, variation which is experienced both in the AAER and the
FS.
However, are these phenomena happening just in these 6 portfolios or are some
trends caused by intrinsic features of the groups and their sets of eligible assets?
Also, in broader approach, what impact do the portfolios’ specific features have in
the value of the MA? In order to answer these questions, we tried to characterize
the distribution of the MA. We already had a Excel workbook that automated the
process of computing the MA and thus we changed it to include two functions:
1. Similarly to the functions that find the Highest Yield and Highest Rating
portfolios, we build a function that finds a random portfolio.
2. To be able to capture the distribution of the measure, we would have to com-
pute it for many random portfolios. Thus, we build an iterative function that
runs the MA’s computation process for 1000 random Matching portfolios and
stores that information in a Excel spreadsheet.
Figure 4.1: General distribution of the MA, in basis points, per group
Figure 4.1 contains box-plots with the general MA’s distribution for all the three
groups. The data clearly supports the trends previously observed. Moreover, Tables
Chapter 4.Matching Adjustment Application 27
4.4 and 4.5 present the average values and the standard error of all major portfolio’
features and all relevant figures influencing the MA’s calculation.
Group MeasureAverage
YieldAverage
CQSAverageDuration
% Invested inGovernment
% ofMismatch
BD Average 2.38 3.64 19.72 60.85% 0.08%S. Error 0.13 0.22 0.23 7.00% 0.05%
MD Average 2.09 3.59 25.34 54.05% 0.48%S. Error 0.11 0.22 0.22 6.79% 0.35%
SD Average 2.03 3.13 19.29 48.59% 3.69%S. Error 0.13 0.23 0.26 7.37% 1.87%
Table 4.4: MA distribution’s general features
Group Measure AAER - LAER - FS = MA
BD Average 194.75 - 170.13 - 2.12 = 22.50S. Error 13.54 0.00 0.71 13.95
MD Average 170.28 - 196.63 - 3.12 = -29.46S. Error 13.20 0.00 0.91 13.91
SD Average 179.74 - 166.03 - 2.09 = 11.63S. Error 18.93 0.00 0.77 19.30
Table 4.5: MA distribution and relevant spreads, in basis points
Analyzing all the data in detail, we conclude the following:
• The MD group shows a consistently lower MA when compared with the other
groups. This is caused by the particular features of the portfolio, particularly,
its longer maturities and low yields.
• The SD group experienced the higher MA’s standard error, which corroborates
the initial suggestion of higher variation. The same happens for almost all its
portfolio’s features.
• The AAER expresses the assets portfolios’ yield. Thus, higher average yields
lead to higher AAER’s. There is an exception, though. Since this spread is
computed based on the liabilities cash-flows, any mismatch between assets and
liabilities may cause unexpected fluctuations in the value of the AAER. This
Chapter 4.Matching Adjustment Application 28
was particularly the case of the Highest Yield portfolios of the groups BD and
MD.
• Since the method doesn’t change the value of liabilities, the LAER is constant.
Thus, the initial trend is maintained, where the average SD portfolio has the
lowest LAER, followed by the BD’s and finally the MD’s. This spread corre-
sponds to the single rate which, when discounting the corresponding liabilities,
is equivalent to the risk-free rates published by EIOPA. Thus, the LAER can
be acknowledged as the ”liabilities’ maturity-weighted” average of the risk-free
rates, which implies that longer tails in the liabilities cash-flows lead to higher
LAER’s (the higher the maturity date, the higher is risk-free rate). Indeed,
this is consistent with the work developed in Chapter 3, in which the SD group
showed the smallest tail, followed by the BD group and the MD group.
• In terms of the FS, the MD group has the highest average, followed by the BD
group and then the SD group (although the BD and SD groups are very close
and, in some cases, the BD group has lower FS’s). The measure is complex
and many portfolio’s features have impact on its value. Firstly, assets with
longer maturities have higher risks of default and downgrade, and thus higher
FS’s. Secondly, assets with high credit worthiness are considered less risky
and so have generally lower FS’s (lower CQS leads to lower FS’s). The same
happens with the investment in central Government and central banks debt -
is is considered less risky and has generally lower FS, which helps explain why
the BD group has a lower average FS than the MD’s and, at the same time, a
higher average CQS.
• The percentage of mismatch remains acceptable. Furthermore, the average SD
portfolio has the lowest mismatch, followed by the MD’s and finally the SD’s.
This is due to the dimension of the liabilities. Actually, once the percentage
of mismatch is computed using the total value of liabilities, the lower it is, the
higher is the impact of the rounding done in the number of units invested in
each asset, which increases the percentage of mismatch.
Chapter 5
Quantitative Impact
5.1 Best Estimate
Under the Solvency II regime, all undertakings have to hold enough assets to cover
all Technical Provisions, and thus, changes in the value of Technical Provisions have
implications in the undertakings’ solvency position. This is the relevance of the BE,
which is the part of Technical Provisions that corresponds to the expected present
value of all future cash-flows. The cash-flows include premiums, benefits, expenses,
and so on. However, in our case, we only considered benefits because the D&D
liabilities portfolios do not include material expenses nor premium payments. Thus,
in order to compute the BE, we just needed the liabilities cash-flows and the relevant
interest rate term structure.
In Chapter 3, we computed the liabilities’ yearly cash-flows since 2015 and, in Chap-
ter 4, we computed the value of the MA’s for all six portfolios. Then, we computed
the adjusted interest rate as the MA plus the relevant risk-free interest rate term
structure published by EIOPA. For each group of undertakings, Table 5.1 describes
the values of the BE for the two portfolios constructed in chapter 4 and for a situation
without the application of the MA, which is identified as the baseline.
29
Chapter 5.Quantitative Impact 30
Group PortfolioMA
(in basispoints)
Best Estimate(in million
euros)
RelativeImpact
BD Baseline 0.00 512.50 0.00%H. Yield 101.27 427.70 -16.55%H. Rating 1.79 510.79 -0.33%
MD Baseline 0.00 79.84 0.00%H. Yield 51.08 71.13 -10.91%H. Rating -34.34 86.56 8.42%
SD Baseline 0.00 13.17 0.00%H. Yield 106.23 10.90 -17.19%H. Rating -13.51 13.50 2.55%
Table 5.1: Quantitative impact in the Best Estimate
As expected, there’s a direct relation between the MA and the impact of its appli-
cation in the BE - the higher the MA, the lower the BE and the more favorable it
is for an undertaking. In fact, the higher the MA, the higher is the interest rate
term structure and consequently the lower is the liabilities’ present value. Also, a
negative value in the MA generates an increase in the BE, which is disadvantageous
for the undertaking.
The more favorable impact can be seen in the Highest Yield portfolios, with a
decrease in the BE varying between 17,19% and 10,91%, while in the Highest Rating
portfolios there’s a variation from -0,33% to 8,42%. This fact seams to indicate
that, for the purpose of building the Matching portfolios, choosing profitability over
security would be preferable. Nevertheless, the impact on the SCR may revert this
scenario.
The group of undertakings where the application of the measure is more favorable
in terms of the BE is the BD group, with all portfolios showing a decrease in the
BE. The less favorable is the MD group, with all portfolios increasing the value of
the BE unless the Highest Yield portfolio.
Chapter 5.Quantitative Impact 31
5.2 Solvency Capital Requirement
The SCR is the measure corresponding to the capital needed to sustain a shock
corresponding to the 99.5% value-at-risk for an one year time horizon. Undertakings
have to compute their SCR at least once a year and should hold enough Basic Own
Funds (BOF) to fully cover that value. Therefore, changes in the SCR’s value have
implications in the undertakings’ solvency position.
In order to compute the SCR, undertakings are given two options - either use the
SCR standard formula set out in the Directive or use an internal model, which is a
formula built by the undertaking and approved by the supervisory authority. For
the purpose of the quantitative impact analysis, we chose to use the SCR standard
formula.
In the standard formula, whose full structure is represented in appendix C, the SCR
is computed as the sum of 3 modules, namely the adjustment for the loss absorbing
capacity of Technical Provisions and deferred taxes, the operational risk and the
Basic SCR. The first two have a specific formula set out in the Delegated Acts,
while the Basic SCR is itself divided into 6 modules, some of them divided into
sub-modules. Each module/sub-module represents a specific risk an undertaking
may be exposed to and its calculation is based on the impact of a stressed scenario
in the BOF. All stressed scenarios are set as specific shocks calibrated at European
level to reflect the aggregated 99.5% value-at-risk for the one year time horizon. In
particular, the basic SCR is computed through the following steps:
1. The whole balance sheet is valued in normal conditions, using proper methods
based on market consistent principles. The BOF are computed as the excess of
assets over liabilities, taking into account the eligible subordinated liabilities.
2. Given a specific module/sub-module, a shock relating the risk considered in
that module/sub-module is applied and the balance sheet is revalued. The
Chapter 5.Quantitative Impact 32
stressed BOF item is again computed. During the revaluation, the Risk Mar-
gin is assumed not to change and the loss-absorbing capacity of Technical
Provisions and deferred taxes is not accounted for.
3. The capital charge for that module/sub-module is computed as the difference
between the normal conditions’ and the stressed BOF. Negative variations in
BOF are recognized as positive capital charges. If the variation in BOF has a
positive sign, then the capital charge is set to zero.
4. After the capital charges for all shocks are computed, they are aggregated
assuming Normal distributions and using linear correlation matrices published
in the Delegated Acts (European Commission, 2014).
The SCR is a very complex and sophisticated formula, built to quantify all the risks
to which an undertaking is exposed to. It is not the purpose of this thesis to go into
much detail on this topic. Therefore, we tried to simplify the computation as much
as possible, always taking into account that the end-goal is to compute the impact
of applying the MA to a general Portuguese undertaking.
5.2.1 SCR in Matching Adjustment Portfolios
All six portfolios presented in chapter 4 are similar in terms of risk exposure for
all three groups of undertakings. On the market risks’ side, because the Match-
ing portfolios have a very special composition of eligible assets, the only material
risk exposures are the interest rate, spread and concentration risks. On the side of
the underwriting risks, because the liabilities only include D&D pensions, the only
material risk exposure is the similar-to-life techniques (SLT) health underwriting
risk, more specifically, the corresponding longevity and revision risks. Both counter-
party default’s and intangible asset’s risks are considered non-existent. Finally, we
included the adjustment for deferred taxes module and the operational risk module.
Chapter 5.Quantitative Impact 33
As before, we built an Excel workbook which includes VBA programmed functions
in order to make the calculations as automated as possible. We also computed
two SCR’s for each portfolio - one considering application of the MA and another
not considering it, which we called the baseline portfolio. Table 5.2 resumes the
computed values and the relative impact of the MA’s application.
Group PortfolioMA
(in basispoints)
SCR(in million
euros)
Baseline SCR(in million
euros)
RelativeImpact
BD H. Yield 101.27 60.40 107.83 -43.98%H. Rating 1.79 61.28 75.56 -18.90%
MD H. Yield 51.08 17.08 27.31 -37.49%H. Rating -34.34 23.01 20.19 13.98%
SD H. Yield 106.23 1.53 3.46 -55.68%H. Rating -13.51 2.86 2.92 -1.91%
Table 5.2: SCR for MA Portfolios and corresponding baselines
The results indicate that there’s a benefit in terms of SCR when applying the MA,
since almost every portfolio had a reduction in their SCR’s, even the SD group’s
Highest Rating portfolio, which has a negative MA. This impact is more accentu-
ated in the Highest Yield portfolios and indicates that higher values of MA lead to
more beneficial impacts. Moreover, when computing risk modules based on shock
scenarios, we must not take into account the MA and thus the impact here is null.
The exception, and reason behind the reduction experienced, is the market spread
risk sub-module, which includes a specific adjustment in case of applying the MA.
This is logical and expected, since, as discussed earlier, the criteria to build eligible
MA portfolios mitigates the risk of losses due to spread fluctuations.
5.2.2 Approximation for an undertaking applying the MA
In Solvency II, when computing the SCR, undertakings applying the MA have to
treat the corresponding portfolios as an independent undertaking. Until now, we’ve
Chapter 5.Quantitative Impact 34
analyzed the impact of applying the MA only on those portfolios. However, no
undertaking deals exclusively on D&D pensions and thus, to fully access this impact,
we need to consider the rest of the undertaking’s business. Specially in terms of the
SCR, because there’s a loss of diversification benefits, analyzing the entire business
may lead to different conclusions.
In order to perform this analysis, we used data from the quantitative report for
the preparatory phase of Solvency II carried out by ASF in the Portuguese insur-
ance market, which included the value and computation of each undertaking’s SCR.
However, due to many problems with the reported data, we had to exclude some
companies from the initial sample set. In particular, for both MD and SD groups,
we ended with just one undertaking, which is considered to represent well the group
and has an acceptable report quality. Then, for the BD group, we considered all
undertakings and thus, we aggregated each specific SCR’s sub-module by taking the
weighted average according with the market shares. The goal here was to adapt the
sub-modules’ reported values to simulate the application of the MA in an undertak-
ing and, for that, we had to make some assumptions, namely:
• In a real situation, an undertaking intending to apply the MA and who does
not have an asset portfolio which replicates the correspondent liabilities, has
to buy new assets that meet all the requirements. Thus, to simulate this trade,
we chose EU government bonds with fixed cash-flows from the undertaking’s
original assets portfolio, which, when sold at their quoted market value in the
end of 2014, were enough to buy the correspondent Matching portfolio. In
terms of the rest of the business’ SCR, this sale only affects the interest rate
and the concentration sub-modules, although in the case of the concentration
sub-module, the sale would be likely immaterial and reduce the capital charge.
Therefore, we only considered the impact of the sale in the interest rate risk
sub-module.
• The lines of business incorporated in the SLT health underwriting risk sub-
module are mainly the D&D pensions and the Portuguese specific life-long
Chapter 5.Quantitative Impact 35
assistance. However, in the Portuguese market, the second is so much less
significant than the D&D pensions that we can consider it immaterial. Thus,
in terms of the rest of the business’ SCR, we assume that removing this sub-
module would be same as removing the exact underwriting risks related with
the D&D pensions’ liabilities portfolio.
And finally, the method used was the following:
• Collect the value of each SCR Standard Formula’s module or sub-module;
• Annul the SLT Health sub-module and subtract the amount of the interest
rate risk sub-module corresponding to assets chosen to be sold in exchange for
the Matching portfolio;
• Aggregate the resulting modules/sub-modules according with the Standard
Formula’s rules, obtaining the henceforth called SCR of the rest of the business;
• Add the SCR of the rest of the business with the SCR of the Matching portfolio,
obtaining the final SCR with the application of the MA.
5.3 Final results and remarks
5.3.1 General impact of the Matching Adjustment
In order to measure the impact of applying the MA in an undertaking, we chose 3
metrics relating with an undertakings’ available assets, risk exposure and solvency
position. Respectively, they are:
Basic Own Funds - Using again the quantitative report for the preparatory phase
of Solvency II, we extracted the amount of BOF eligible to cover the SCR of
each undertaking considered in the analysis of the SCR (i.e., all companies
from the BD group and one company for both the MD and SD groups). It
corresponded to each group’s initial value of BOF eligible to cover the SCR.
Chapter 5.Quantitative Impact 36
Then, to compute this measure after the application of the MA, we subtracted
the impact observed in the BE from Table 5.1, adjusted to take into account
the increase in deferred taxes. The formula was:
BOFafterMA = BOFinitial − (BEafterMA −BEinitial)(1− 0.23) (5.1)
SCR - All the computations were already done here. We used the initial SCR of
each group and their final SCR including the application MA.
Solvency Ratio - This ratio measures the solvency of an undertaking, that is, its
ability of meeting future liabilities. It is computed as the ratio of BOF to cover
the SCR over the SCR itself.
The relative impact on the three metrics is presented in Table 5.3.
Group PortfolioMA
(basis points)BOF SCR
SolvencyRatio
BD H. Yield 101.27 14.21% 12.01% 1.97%H. Rating 1.79 0.29% 6.85% -6.14%
MD H. Yield 51.08 5.64% 17.42% -10.03%H. Rating -34.34 -4.35% 23.35% -22.46%
SD H. Yield 106.23 14.84% 12.64% 1.96%H. Rating -13.51 -2.20% 21.71% -19.65%
Table 5.3: Relative impact on solvency of the Matching Adjustment
And finally, the main conclusions are the following:
• First of all, it is clear that the higher the value of the MA, the higher is the
value of BOF, which is favorable for an undertaking. In fact, with the MA,
the Best Estimate decreases, and thus, the undertaking is considered to have a
lower expected value of liabilities, which increases the undertaking’s available
capital. However, due to a loss in deferred taxes, this benefit is slightly reduced.
Chapter 5.Quantitative Impact 37
• In terms of the SCR, the groups were always penalized for applying the MA.
Despite the reduction in the market’s spread risk submodule, the loss of di-
versification benefits when applying the MA caused the SCR to increase in
every portfolio, which implies an increase in the assets allocated to sustain
unexpected losses.
• On one hand, the MA was beneficial because it increased the undertaking’s
available capital. Yet, on the other hand, the MA was disadvantageous because
it increased the capital needed to sustain unexpected losses, which decreased
the accessible capital an undertaking has to expand and/or create new busi-
ness. This duality is the reason why we chose to analyze the solvency ratio -
since it balances both BOF and SCR, it represents the general impact in an
undertaking. According to this metric, there are only two portfolios where it
was advantageous to apply the MA. They are the BD and SD groups’ Highest
Yield portfolios. In these portfolios, the relative increase in own funds was
higher than the relative increase in the SCR, which led to an improvement in
the solvency position of the undertaking.
5.3.2 General impact of the Volatility Adjustment
The last goal proposed to the internship was to compare the MA with other LTG
measures, in particular, the Volatility Adjustment (VA). This measure was built by
EIOPA to mitigate pro-cyclical investment behaviors when markets deteriorate due
to wide bonds’ spreads and bonds’ low liquidity. The VA is a risk-corrected spread
which is added to the risk-free interest rate term structure used to discount Technical
Provisions and is calculated based on reference portfolios for each currency.
It corresponds to the difference between the rates obtained from a reference port-
folio and the basic risk-free rates. All computations are performed periodically and
published by EIOPA.
Chapter 5.Quantitative Impact 38
Similarly to the MA, we assessed the impact of applying the VA on three items - the
BOF, the SCR and the Solvency ratio. The results are summarized in Table 5.4.
Group BOF SCR Solvency Ratio
BD 2.26% 0.00% 2.26%MD 1.94% 0.00% 1.94%SD 2.56% 0.00% 2.56%
Table 5.4: Relative impact on solvency of the Volatility Adjustment
In terms of the BOF, the impact was computed using formula 5.1, where the index
”MA” is replaced by ”VA”. Similarly to the MA, because there’s a rise in the
interest rates used to discount Technical Provisions, when applying the VA, the BE
decreases, which increases the value of Own Funds and results in a favorable impact
for the undertaking. Secondly, in terms of the SCR, the measure has no impact on
the standard formula’s modules computed by stressed scenarios and thus, for every
group of undertakings, the impact observed in the SCR was null and the impact
experienced in the Solvency Ratio is same as the BOF’s.
Finally, comparing these results with the ones on Table 5.3, we conclude that for
all groups, the VA was more favorable than the MA, since it caused higher positive
impacts. Nevertheless, the VA is very dependent on market conditions and it is
only relevant for insurers in cases of market instability. Thus, with different market
conditions, even if an undertaking is allowed to apply the VA, it is not certain that
the VA will always be more favorable then the MA to undertakings.
Chapter 6
Conclusion
The work developed during my internship at ASF was aimed to assess the impact of
applying the MA in a specific line of business in Portugal - the Workers’ Compensa-
tion insurance. In particular, we focused on the Death and Disability pensions due
to their features, which are well suited for applying the MA.
In order to better represent the market, we decided to divide all undertakings with
business in WC insurance and supervised by ASF in 3 groups, according with their
market shares – the BD group represented 81% of the market, the MD group had
17% of the market and the SD group constituted 2% of the market. In fact, we
considered the market shares weighted average of each group’s undertakings, creating
thus notional undertakings that were representative of the respective groups. Then,
using the 3 groups, we did a series of computations and analysis based on real data
reported to ASF. Our main conclusions were the following:
• As expected, the current WC insurance market is not yet prepared to apply
the MA since the assets portfolios don’t fulfill the mandatory criteria and the
liabilities and assets cash-flows are not matched at all times (assets cash-flows
are higher the the first maturity years and have much shorter maturities then
the liabilities cash-flows).
39
Chapter 6.Conclusion 40
• The MD group presents generally lower values of MA than the other two
groups, mainly due to its liabilities’ longer tail and its assets’ yields. Also,
the BD and SD groups are similar in terms of main portfolio features and MA
value. However, the SD group presents a higher variability and higher value
of mismatch between assets and liabilities. This is caused by the difference of
their liabilities’ dimension.
• In terms of Technical Provisions, the Best Estimate is the component that
experiences the most relevant impact. In fact, there’s a direct relation between
the measure and the MA - the higher the MA the lower is the relative value
of the Best Estimate and the more beneficial it is for the undertaking.
• Own Funds are directly influenced by changes in Technical Provisions. Ap-
proximately, the BOF are simply the excess of assets over liabilities and, since
a decrease in the Best Estimate causes a decrease in liabilities, the higher
the value of the MA, the higher are the BOF. Therefore, in terms of BOF,
even after the adjustments due to the loss in deferred taxes, we observed an
improvement when applying the MA.
• In the case of the SCR, there are two different situations. Firstly, when we
only consider the SCR of the D&D related portfolios, due to the decrease in
the spread risk sub-module, the application of the MA reduces the value of
the SCR, which is beneficial for an undertaking. However, when we look at
the global impact in an undertaking, the application of the MA increases the
value of the SCR, because the benefit in the spread risk sub-module isn’t high
enough to offset the loss of diversification benefits. Thus, in terms of the SCR,
applying the MA is disadvantageous for an undertaking.
• In order to measure the global impact of applying the MA in an undertak-
ing, we looked at the Solvency ratio, which considers both SCR and BOF.
According with this metric, only two portfolios would lead to a benefit - the
BD’s and SD’s Highest Yield portfolios. In all the others, the value of the
MA wasn’t high enough to lead to a relative increase in BOF higher than
the one observed in the SCR. This fact seems to imply that, when building
Chapter 6.Conclusion 41
the Matching portfolio, an undertaking should choose profitability over credit
quality, which appears to be counter-intuitive. Nevertheless, the use of the
MA in each portfolio has to be approved by the supervisory authority and
thus, in practice, an undertaking will have to guarantee a certain degree of
safety in credit worthiness. Also, the measure appears to be more beneficial
to either undertakings with a large business or to undertakings with an very
representative WC business, since, in both cases, the loss in diversification
benefits would be relatively lower.
• Finally, in the current market conditions, the impact of applying the VA is
more favorable in every group of undertakings than the impact of applying the
MA. However, this result doesn’t mean that the VA is always more favorable
than the MA, since it is very dependent on market conditions and can be only
applied in specific cases where the markets are unstable.
Clearly, applying the MA can be advantageous to an undertaking. However, due
to the fact that, once applying the measure, an undertaking cannot revert back to
not applying it, it’s essential to do a complete study of the MA’s behavior in the
specific undertaking under various market scenarios. Also, the results indicate that
in undertakings with a similar profile to the MD group it wouldn’t be advantageous
to apply the MA. Nonetheless, the MA is so tailor-made that small differences in
the undertaking’s business and in the D&D portfolios may cause huge differences
in the final impact. Therefore, an interesting development of this work would be to
focus on a specific Portuguese undertaking and do a thorough analysis.
Finally, by implementing the process in Excel and VBA, we hoped to contribute to
the analysis on the LTG problem and to make the MA easier to implement in the
Portuguese market. An undertaking can base their computation on this work, built
upon it and use the automated Excel workbook to test various market situations.
Also, we chose very specific investment strategies for building the Matching portfolios
and there is no reason why an undertaking shouldn’t experiment different strategies,
which would be a great development of this work.
Appendix A
Technical Specification for the
Liabilities Portfolio
A.1 Annual aggravation rates
Table A.1 contains the percentage of revisions on D&D pensions observed in the
data corresponding to years 2012, 2013 and 2014, contains the average percentage of
the pension’s aggravation, in case a revision has occurred, and the final aggravation
rate, which is give by the product of the two.
Percentageof revisions
Average Aggravation Aggravation rate
Year 2012 2013 2014 2012 2013 2014 2012 2013 2014
DisabilityRevision
2,89% 3,49% 2,55% 54,19% 54,55% 22,09% 1,57% 1,90% 0,56%
State changeto Definitive
2,29% 2,46% 2,57% 10,08% 21,39% 20,83% 0,23% 0,53% 0,54%
Total5,18% 5,95% 5,12% 64,27% 75,94% 42,92% 3,33% 4,52% 2,20%
Table A.1: Aggravation Analysis
42
Appendix B
Matching Portfolios
43
Appendix B. Matching Portfolios 44
BD Group
Highest Yield portfolio Highest Rating portfolio
ISIN Yield Rating Units ISIN Yield Rating Units
XS0503253345 1,617 NR 35 AT0000386198 -0,099 AA+ 2313PTRELYOE0002 1,116 NR 2270 XS0457688215 0,454 AAA 2704
XS0276455937 7,462 BB+ 2598 XS0282445336 0,288 AA- 3115XS0716979249 5,601 BB+ 3067 XS0695461458 0,344 AA- 3550
PTCFPAOM0002 2,452 NR 83 XS0801654558 0,344 AAA 3978XS0906946008 6,721 BB+ 4056 XS0478074924 0,582 AA- 4348XS0919581982 8,617 BB+ 4553 XS1130487868 0,608 AAA 4803XS0716979595 6,447 BB+ 5067 XS1135334800 0,825 AA+ 5101XS0835890350 6,092 BB+ 5717 FR0011486067 0,584 AA+ 5379174
PTOTEQOE0015 2,654 NR 563223397 XS1121229402 1,442 A- 5780XS0221854200 4,562 BB 7038 ES0413211071 1,061 A+ 61IT0001086567 2,188 NR 2576 ES00000123C7 1,933 BBB 6471XS0862347852 2,562 BBB 7742 FR0011370386 1,441 AAA 71ES00000124C5 2,235 BBB 7761 ES00000124C5 2,235 BBB 7416ES0000011868 2,239 BBB 833014315 ES0000011868 2,239 BBB 796748662
PTOTEROE0014 3,424 NR 896146075 PTOTEROE0014 3,424 NR 857704483IT0001444378 2,612 NR 9409 XS0129547948 2,225 AAA 182ES0000012411 2,4 BBB 10037 ES0000012411 2,4 BBB 9631FR0010014845 5,15 BB- 13785 FR0010014845 5,15 BB- 13230IT0003535157 2,904 NR 11288 IT0003535157 2,904 NR 10833XS0223484345 4,308 BBB- 11818 BE0002436112 1,508 AAA 115ES0001350125 3,823 BB+ 122 BE0002440155 1,478 AAA 118
PTOTE5OE0007 3,678 NR 1257742228 ES0000012932 2,651 BBB 11824XS1069178686 2,928 BBB 148 BE0002431063 1,594 AAA 124IT0004286966 3,105 NR 13219 IT0004286966 3,105 NR 12322IT0004976764 3,687 NR 1364837521 XS0884635524 1,503 AAA 12707XS0442127063 5,827 NR 19612 XS0881486657 2,18 AA+ 132IT0004976806 3,741 NR 1307135988 XS0752034206 1,521 AAA 12823XS0208362961 6,006 NR 19049556 LU0962396452 1,616 AAA 129IT0004976830 3,657 NR 1296132730 FR0012039931 1,722 AA 129
AT0000A0CJE9 2,576 NR 1890555314 XS1107247725 1,577 AAA 12866AT0000A0CJF6 2,578 NR 1824334921 DE0001102341 1,384 AAA 1265161677XS0295018070 4,952 BB+ 17502 XS1061430051 2,045 AA 133XS0181673798 4,193 NR 11741 XS0352558422 0,365 AA 127
AT0000A0CJJ8 2,578 NR 1784321610 DE0002942448 1,582 AAA 25130AT0000A0CJK6 2,578 NR 1702521919 DE000NRW0DB8 1,538 AA 11623AT0000A0CJL4 2,578 NR 1617699310 AT0000A0CJL4 2,578 NR 1716977641XS1055501974 5,619 NR 15305 XS1055501974 5,619 NR 16298XS0970703772 3,605 NR 96 XS0877035765 1,743 AA+ 107
AT0000A0CJP5 2,578 NR 1423555861 AT0000A0CJP5 2,578 NR 1450044008XS0214965963 5,214 BB+ 133 FR0010292169 2,002 AA+ 9428
AT0000A0CJR1 2,578 NR 1308995536 AT0000A0CJR1 2,578 NR 1318198906XS0292467775 11,065 CCC 8 AT0000A0CJS9 2,578 NR 1225358072
AT0000A0CJT7 2,578 NR 1147791620 AT0000A0CJT7 2,578 NR 1133028239AT0000A0CJU5 2,578 NR 1055990280 AT0000A0CJU5 2,578 NR 1041226898XS0515753183 3,347 NR 6408 FR0010870956 2,005 AA+ 6309528
AT0000A0CJW1 2,578 NR 923148208 AT0000A0CJW1 2,578 NR 899590778AT0000A0CJX9 2,578 NR 836300289 XS0757586267 2,256 AA 58XS0936805612 3,649 NR 50 XS0939098363 2,089 AA 54ES00000126D8 3,162 BBB 4695 ES00000126D8 3,162 BBB 4695
AT0000A0CK09 2,578 NR 658148790 AT0000A0CK09 2,578 NR 658148790AT0000A0CK17 2,578 NR 584984876 AT0000A0CK17 2,578 NR 584984876AT0000A0CK25 2,578 NR 516058512 AT0000A0CK25 2,578 NR 516058512AT0000A0CK33 2,578 NR 451763098 AT0000A0CK33 2,578 NR 451763098AT0000A0CK41 2,578 NR 392180463 AT0000A0CK41 2,578 NR 392180463AT0000510102 2,578 NR 337578676 AT0000510102 2,578 NR 337578676
AT0000A0CK66 2,578 NR 288167526 AT0000A0CK66 2,578 NR 288167526AT0000A04ER3 2,578 NR 243586455 AT0000A04ER3 2,578 NR 243586455AT0000A0CK82 2,578 NR 203941062 AT0000A0CK82 2,578 NR 203941062AT0000A0BNE3 2,578 NR 801207982 AT0000A0BNE3 2,578 NR 801207982
Appendix B. Matching Portfolios 45
MD Group
Highest Yield portfolio Highest Rating portfolio
ISIN Yield Rating Units ISIN Yield Rating Units
ES0413900103 0,236 AA 0 ES0413900103 0,236 AA 0FR0011512417 1,647 NR 0 FR0010717785 0,055 AAA 0
PTBITIOM0057 2,656 BB- 0 XS0282588952 0,258 AA+ 0XS0982711631 5,441 BB+ 9 ES0413900327 0,343 AA 0XS0184373925 2,294 BB+ 1 XS0193640629 0,275 AAA 0XS0970695572 2,062 BB+ 120 XS0237323943 0,618 AA- 68
PTOTEYOE0007 1,943 NR 16923543 XS0526606537 0,59 AA- 123IT0004848831 1,597 NR 226 FR0011337880 0,518 AA+ 177601XS0971213201 3,516 BB+ 321 DE000A0PM5F0 0,483 AAA 235
PTCPEJOM0004 3,524 NR 10 FR0011521319 0,693 AA 300XS0994991411 1,34 AA- 4 XS0994991411 1,34 AA- 4ES0413900129 1,169 AA 5 ES0413900129 1,169 AA 4XS0306646042 1,993 BBB 620 XS0288429532 1,358 AA- 10IT0004889033 2,46 NR 684 XS0093667334 1,25 AAA 74246ES0000011868 2,239 BBB 77925871 FR0000571218 1,256 AA+ 689875XS0506861243 4,248 BB 889 XS0506264315 1,75 AAA 16XS0137905153 4,745 BB 12 XS0129547948 2,225 AAA 18ES0000012965 2,705 NR 1118099 FR0000187635 1,472 AA+ 997687IT0003256820 2,757 NR 1180 XS0162869076 2,504 BBB 1225XS0191352847 9,571 CCC 1307 BE0002421932 1,572 AAA 13BE0000304130 1,611 AA- 143145990 BE0000304130 1,611 AA- 133547590ES0414950644 2,322 A 32 ES0414950644 2,322 A 30
PTOTE5OE0007 3,678 NR 167321149 AT0000A04967 1,343 AA+ 1568FR0010371401 1,752 AA+ 1788440 FR0010371401 1,752 AA+ 1679981DE0001135325 1,295 NR 190348658 DE0001135325 1,295 NR 179068929IT0004976764 3,687 NR 202252522 XS0884635524 1,503 AAA 1905XS0442127063 5,827 NR 3010 XS0881486657 2,18 AA+ 21IT0004976806 3,741 NR 207817797 XS0752034206 1,521 AAA 2066XS0208362961 6,006 NR 3134405 LU0962396452 1,616 AAA 22IT0004976830 3,657 NR 220831904 FR0012039931 1,722 AA 22
AT0000A0CJE9 2,576 NR 333306444 XS1107247725 1,577 AAA 2296AT0000A0CJF6 2,578 NR 332189244 DE0001102341 1,384 AAA 232845293XS0295018070 4,952 BB+ 3302 XS1061430051 2,045 AA 25XS0181673798 4,193 NR 2293 XS0352558422 0,365 AA 25
AT0000A0CJJ8 2,578 NR 360003205 DE0002942448 1,582 AAA 5100AT0000A0CJK6 2,578 NR 355777961 DE000NRW0DB8 1,538 AA 2441AT0000A0CJL4 2,578 NR 349521904 AT0000A0CJL4 2,578 NR 372121581XS1055501974 5,619 NR 3418 XS1055501974 5,619 NR 3644XS0970703772 3,605 NR 22 XS0877035765 1,743 AA+ 25
AT0000A0CJP5 2,578 NR 341443461 AT0000A0CJP5 2,578 NR 347184276XS0214965963 5,214 BB+ 33 FR0010292169 2,002 AA+ 2335
AT0000A0CJR1 2,578 NR 337733942 AT0000A0CJR1 2,578 NR 339151653XS0292467775 11,065 CCC 2 AT0000A0CJS9 2,578 NR 327010968
AT0000A0CJT7 2,578 NR 319519828 AT0000A0CJT7 2,578 NR 314521075AT0000A0CJU5 2,578 NR 307277312 AT0000A0CJU5 2,578 NR 302278560XS0515753183 3,347 NR 1955 FR0010870956 2,005 AA+ 1921983
AT0000A0CJW1 2,578 NR 294288842 AT0000A0CJW1 2,578 NR 286586955AT0000A0CJX9 2,578 NR 279917388 XS0757586267 2,256 AA 19XS0936805612 3,649 NR 18 XS0939098363 2,089 AA 19ES00000126D8 3,162 BBB 1760 ES00000126D8 3,162 BBB 1760
AT0000A0CK09 2,578 NR 260466510 AT0000A0CK09 2,578 NR 260466510AT0000A0CK17 2,578 NR 246089124 AT0000A0CK17 2,578 NR 246089124AT0000A0CK25 2,578 NR 231955906 AT0000A0CK25 2,578 NR 231955906AT0000A0CK33 2,578 NR 216291338 AT0000A0CK33 2,578 NR 216291338AT0000A0CK41 2,578 NR 201851342 AT0000A0CK41 2,578 NR 201851342AT0000510102 2,578 NR 186481847 AT0000510102 2,578 NR 186481847
AT0000A0CK66 2,578 NR 171754251 AT0000A0CK66 2,578 NR 171754251AT0000A04ER3 2,578 NR 157225433 AT0000A04ER3 2,578 NR 157225433AT0000A0CK82 2,578 NR 141622055 AT0000A0CK82 2,578 NR 141622055AT0000A0BNE3 2,578 NR 976229911 AT0000A0BNE3 2,578 NR 976229911
Appendix B. Matching Portfolios 46
SD Group
Highest Yield portfolio Highest Rating portfolio
ISIN Yield Rating Units ISIN Yield Rating Units
PTBLMXOM0019 1,821 NR 1 XS0541454467 0,169 AA- 79PTPTCYOM0008 3,696 NR 66 XS0844535442 0,084 AA 88
XS0215828913 2,636 BB 75 FR0011374198 0,151 AAA 1XS0630463965 1,845 BB+ 85 XS0986254455 0,313 AAA 104XS1017435782 3,26 BBB- 91 DE000A1AKHB8 0,363 AA- 2XS0970695572 2,062 BB+ 108 FR0011431014 0,324 AAA 1
PTGALJOE0008 3,217 NR 1 XS0693812355 0,688 AA- 129XS0486101024 3,012 BB+ 3 XS0813400305 0,807 AA- 139XS0835890350 6,092 BB+ 146 FR0011470764 0,715 AAA 2FI4000079041 0,65 AA+ 146 FI4000079041 0,65 AA+ 161XS0982711714 6,15 BB+ 179 IE00B4TV0D44 1,387 A- 17199933XS1057659838 3,074 BBB+ 182 XS1057659838 3,074 BBB+ 199FR0011225143 1,594 A+ 2 FR0011225143 1,594 A+ 2ES0000095606 4,123 BB+ 3455 FR0010970863 1,611 AAA 5XS1120892507 2,264 BBB 2 XS1120892507 2,264 BBB 3IE00BJ38CR43 1,809 A- 23846824 IE00BJ38CR43 1,809 A- 25474122XS0137905153 4,745 BB 3 XS0129547948 2,225 AAA 5XS0522079911 4,064 BB+ 5 FR0011349109 1,86 AAA 3FR0010014845 5,15 BB- 378 DE0003271482 1,741 AAA 30749755BE0000333428 1,61 NR 31456617 BE0000333428 1,61 NR 30833051XS0223484345 4,308 BBB- 325 BE0002436112 1,508 AAA 3ES0001350125 3,823 BB+ 3 BE0002440155 1,478 AAA 3
PTOTE5OE0007 3,678 NR 35446235 ES0000012932 2,651 BBB 341XS1069178686 2,928 BBB 4 BE0002431063 1,594 AAA 4XS0440592748 2,503 BBB+ 9 XS1088825143 1,608 AAA 370IT0004976764 3,687 NR 39485772 XS0884635524 1,503 AAA 365XS0442127063 5,827 NR 569 XS0881486657 2,18 AA+ 4IT0004976806 3,741 NR 37956572 XS0752034206 1,521 AAA 370XS0208362961 6,006 NR 551835 LU0962396452 1,616 AAA 4IT0004976830 3,657 NR 37341738 FR0012039931 1,722 AA 4
AT0000A0CJE9 2,576 NR 53976439 XS1107247725 1,577 AAA 365AT0000A0CJF6 2,578 NR 51446683 DE0001102341 1,384 AAA 35458810XS0295018070 4,952 BB+ 486 XS1061430051 2,045 AA 4XS0181673798 4,193 NR 321 XS0352558422 0,365 AA 3
AT0000A0CJJ8 2,578 NR 47794523 DE0002942448 1,582 AAA 670AT0000A0CJK6 2,578 NR 44635516 DE000NRW0DB8 1,538 AA 303AT0000A0CJL4 2,578 NR 41378341 AT0000A0CJL4 2,578 NR 43776517XS1055501974 5,619 NR 381 XS1055501974 5,619 NR 405XS0970703772 3,605 NR 2 XS0877035765 1,743 AA+ 3
AT0000A0CJP5 2,578 NR 33501187 AT0000A0CJP5 2,578 NR 34137368XS0214965963 5,214 BB+ 3 FR0010292169 2,002 AA+ 216
AT0000A0CJR1 2,578 NR 29084250 AT0000A0CJR1 2,578 NR 29326694XS0292467775 11,065 CCC 0 AT0000A0CJS9 2,578 NR 26504298
AT0000A0CJT7 2,578 NR 24109455 AT0000A0CJT7 2,578 NR 23834356AT0000A0CJU5 2,578 NR 21545718 AT0000A0CJU5 2,578 NR 21270619XS0515753183 3,347 NR 127 FR0010870956 2,005 AA+ 125193
AT0000A0CJW1 2,578 NR 17796944 AT0000A0CJW1 2,578 NR 17348220AT0000A0CJX9 2,578 NR 15777585 XS0757586267 2,256 AA 1XS0936805612 3,649 NR 1 XS0939098363 2,089 AA 1ES00000126D8 3,162 BBB 85 ES00000126D8 3,162 BBB 85
AT0000A0CK09 2,578 NR 11694064 AT0000A0CK09 2,578 NR 11694064AT0000A0CK17 2,578 NR 10257225 AT0000A0CK17 2,578 NR 10257225AT0000A0CK25 2,578 NR 8976319 AT0000A0CK25 2,578 NR 8976319AT0000A0CK33 2,578 NR 7808380 AT0000A0CK33 2,578 NR 7808380AT0000A0CK41 2,578 NR 6722068 AT0000A0CK41 2,578 NR 6722068AT0000510102 2,578 NR 5730975 AT0000510102 2,578 NR 5730975
AT0000A0CK66 2,578 NR 4852405 AT0000A0CK66 2,578 NR 4852405AT0000A04ER3 2,578 NR 4061048 AT0000A04ER3 2,578 NR 4061048AT0000A0CK82 2,578 NR 3379811 AT0000A0CK82 2,578 NR 3379811AT0000A0BNE3 2,578 NR 14478819 AT0000A0BNE3 2,578 NR 14478819
Appendix C
SCR Standard Formula
47
Appendix D
Excel and VBA sample functions
Example of VBA programmed code, used to compute the
Matching Adjustment and the Matching portfolios
’### Computes de MA using solver
Private Sub MAcompute ()
SolverReset
SolverOk SetCell :="$E$5", MaxMinVal :=3, ValueOf :=0, ByChange :="$H$4", Engine :=1,
EngineDesc :="GRG Nonlinear"
SolverSolve True
SolverReset
SolverOk SetCell :="$F$5", MaxMinVal :=3, ValueOf :=0, ByChange :="$J$4", Engine :=1,
EngineDesc :="GRG Nonlinear"
SolverSolve True
End Sub
’### Creates the highest yield portfolio
Sub HighestYield ()
Worksheets("Data Bonds"). Activate
For n = 2015 To 2074
For i = 3 To 700
If year(Worksheets("Data Bonds").Cells(i, 6). Value) = n Then
Set Max = Worksheets("Data Bonds"). Cells(i, 2)
Exit For
End If
48
Appendix D. Excel and VBA sample functions 49
Next i
For x = 3 To 700
If year(Worksheets("Data Bonds").Cells(x, 6). Value) = n Then
If Worksheets("Data Bonds").Cells(x, 5). Value > Max.Offset(, 3) Then
Set Max = Worksheets("Data Bonds"). Cells(x, 2)
End If
End If
Next x
For a = 2 To 16
Worksheets("Portfolio").Cells(n - 2012, a).Value = Max.Offset(, a - 2). Value
Next a
Next n
End Sub
Example of an Excel worksheet, used to compute the Match-
ing Adjustment
Bibliography
C. Alegre. Acidentes de Trabalho e Doencas Profissionais. Edicoes Almedina, 2nd
edition, 11 2011.
S. A. Broverman. Mathematics of investment and credit. Actex Publications, 5th
edition, 2010.
D. Buckham, J. Wahl, and S. Rose. Executive’s Guide to Solvency II. Wiley, 1st
edition, 11 2010.
A. Bar and C. Liao. What do you know about Solvency II?: high-level introduction for
interested parties from non-EU regions. 2013. URL http://media.swissre.
com/documents/Solvency_II_EN.pdf.
Council. Directive 73/239/EEC of 24 July 1973 on the coordination of
laws, regulations and administrative provisions relating to the taking-up
and pursuit of the business of direct insurance other than life assurance,
1973. URL http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=
CELEX:31973L0239&from=EN.
EIOPA. Technical findings on the long-term guarantees assessment. Technical
report, EIOPA, 2014. URL https://eiopa.europa.eu/Publications/QIS/
EIOPA_LTGA_Report_14_June_2013_01.pdf.
EIOPA. Technical document regarding the risk free interest rate term structure,
2015. URL https://eiopa.europa.eu/Publications/Standards/EIOPA_
RFR_Technical_Documentation.pdf.
50
Bibliography 51
European Commission. Delegated Regulation (EU) 2015/35 of 10 October 2014
supplementing Directive 2009/138/EC of the European Parliament and of the
Council on the taking-up and pursuit of the business of Insurance and Reinsur-
ance (Solvency II), 2014. URL http://eur-lex.europa.eu/legal-content/
EN/TXT/PDF/?uri=CELEX:32015R0035&from=EN.
European Parliament and Council. Directive 2002/83/ECC of 5 November
2002 concerning life assurance, 2002. URL http://eur-lex.europa.eu/
legal-content/EN/TXT/PDF/?uri=CELEX:32002L0083&from=EN.
European Parliament and Council. Directive 2009/138/EC of 25 November 2009
on the taking-up and pursuit of the business of Insurance and reinsurance (Sol-
vency II) (Recast), 2009. URL http://eur-lex.europa.eu/legal-content/
EN/TXT/PDF/?uri=CELEX:32009L0138&from=EN.
European Parliament and Council. Directive 2014/51/EU of 16 April 2014 amending
Directives 2003/71/EC and 2009/138/EC and Regulations (EC) No 1060/2009,
(EU) No 1094/2010 and (EU) No 1095/2010 in respect of the powers of the
European Supervisory Authority (European Insurance and Occupational Pen-
sions Authority) and the European Supervisory Authority (European Secu-
rities and Markets Authority) , 2014. URL http://eur-lex.europa.eu/
legal-content/EN/TXT/PDF/?uri=CELEX:32014L0051&from=EN.
Instituto de Seguros de Portugal. Norma Regulamentar N.o 11/2007-R de 26 de
Julho - Sistema de informacao de pensoes de acidentes de trabalho, 2007.
Joint Committee. Draft Implementing Technical Standards on the allocation
of credit assessments of ECAIs to an objective scale of credit qual-
ity steps under Article 109 (a) of Directive 2009/138/EC , 2015. URL
http://www.eba.europa.eu/documents/10180/1006707/JC+CP+2015+001+
%28Joint+CP+on+draft+ITS+on+mapping+of+ECAIs+under+SII%29.pdf.
Bibliography 52
J. L. Raptis. European financial regulation: Esma and the lamfalussy process, the
renewed european legislative process in the field of secutities regulation. The
Columbia journal of European law, 18, 2012. URL http://www.cjel.net/
wp-content/uploads/2012/09/Raptis_61-68.pdf.
P. Sharma. Prudential supervision of insurance undertakings. Technical report,
The London Working Group, 2002. URL http://ec.europa.eu/internal_
market/insurance/docs/solvency/impactassess/annex-c02_en.pdf.